Chapter 6 Prametric Tests

Parametric Tests
Dejen Tesfaw (PhD)

Department of Statistics
AAU , College of Natural and Computational Sciences
Graduate Building-Room N-502
Email: dejentm@gmail.com, or dejen.tesfaw@aau.edu.et
• Parametric tests of hypotheses concerning:
o Variances
o Means
Tests of equality of two population variances
(independent samples)
3
4
5
6
Conclusion:
• There is a significant difference in the
variance of income between defaulters and
non-defaulters at the 5% level of
significance.
• Thus, the fluctuation (dispersion) in the

income level of defaulters is significantly
greater than that of non-defaulters.
7
Practice 1: The following data represents the distance covered by
marathon runners until a physiological stress develops and
whether they used drug or not.
No Distance in Drug No Distance in Drug No Distance in Drug
miles use miles use miles use
1 14.5 no 11 16.9 yes 15.4 no
2 13.4 no 12 12.6 no 22 18.2 no
3 14.8 yes 13 13.4 no 23 14.5 no
4 19.5 yes 14 16.3 yes 24 21 yes
5 14.5 no 15 17.1 yes 25 17.5 yes
6 18.2 yes 16 11.8 no 26 14.2 no
7 16.3 no 17 13.3 yes 27 18.1 yes
8 14.8 no 18 14.5 no 28 13.5 no
9 20.3 yes 19 13.3 No 29 14.1 no
10 18.4 yes 20 17.6 yes 30 13.6 no
Question: Test if there is a significant difference in the mean of distance

covered between drug users and non-users at the 5% level of significance.
Tests concerning the difference between two
means (independent samples)
9
10
11
12
13
14
15
16
17
Note:
• Parametric statistical methods are highly
influenced by extreme values (outliers).
• Thus, we have to check if our conclusion was
affected by extreme values.
• A box plot is a graphical method of detecting

outliers.
• The box plot of the data under consideration is

shown below:
18
• We can see from the box plot that the item on the
445th row is an outlier (extreme value).
19
• After removing this outlier, the independent sample t-test yields
the following result:
20
Remark: The decision as to whether the null hypothesis has
to be rejected or not is based on a p-value that is
produced in statistical software outputs.
Definition: (p-value)
• A p-value is the smallest level of significance or the
smallest value of  for which the null hypothesis is to be
rejected.
For instance:
• If p-value = 0.002, then we can reject the null hypothesis
for all values of  greater than 0.002 (such as  = 0.01,  =
0.05).
• If p-value = 0.03, then we can reject the null hypothesis

for all values of  greater than 0.03 (such as  = 0.05).
However, we can not reject the null hypothesis at  = 0.01.
21
• If p-value = 0.07, then we can not reject
the null hypothesis at both 5% and 1%
levels of significance since 0.07 exceeds
both 0.01 and 0.05.
• Note: In SPSS output, the p-value

corresponds are displayed in the column
under Sig .
22
• In our case, the Levene's test for equality of
variances has a p-value of 0.061 which is
greater than 5%.
• Thus, we do not reject, H0 : 12  12 and

consequently, we refer to the result in the
‘Equal variances assumed’ row.
• The p-value is 0.005 which is less than 5%

(in fact less than 1%).
• Thus, we reject the hypothesis of equality

of means of the two groups.
23
24
25
26
27
Example: This example uses data from SPSS:
property_assess_cs_sample.sav.
• The data is on current value of properties (houses)

categorized by county (location) as Eastern,
Central, Western, Northern and Southern.
• Our aim is to see if there is a significant difference

in the average value of properties across the
various counties.
28
29
30
• Question: which groups of means are different?
• To answer this question, we apply pair-wise

comparison of means.
• Since the equality of variances assumption is

rejected, the appropriate tests are those listed
under ‘Equal Variances Not Assumed’ in SPSS.
• The output of Post Hoc Tests (multiple

comparisons) is shown below:
31
32
• Thus, there is a significant difference in the mean value of properties
across all counties (locations) except those in Eastern and Western.
• The arrangement of counties from the highest mean value of

properties to the smallest is shown below:
• Western and Eastern
• Central
• Southern
• Northern
dejentm@gmail.com 33
• Practice 2: The following table shows the natural killer cell
activity measured for three groups of subjects: those who had
low, medium, and high scores on the social readjustment rating
scale. Question: Test if there is
Low score Moderate score High score a significant difference in
22.2 15.1 10.2 mean natural killer cell
activity among patients
97.8 23.2 11.3
with low, moderate and
29.1 10.5 11.4 high scores on the social
37.0 13.9 5.3 readjustment rating scale
35.8 9.7 14.5 at the 5% level of
44.2 19.0 11.0
significant.
88.0 19.8 13.6
56.0 9.1 33.4
9.3 30.1 25.0
19.9 15.5 27.0
39.5 10.3 36.3
12.8 11.0 17.7
37.4 34
Practice 3: The following table gives the birth weights of 24 randomly taken live
born infants. Infants were classified by the socio-economic status of their
families (i.e., group 1= poor; group 2=average; group 3= comparatively rich). Test
there is a significant difference in the mean value of weights with new born
infants across the socio-economic status of their family.
Group 1 (n = 10) Group 2 (n = 8) Group 3 (n = 6)

2.60 2.50 3.30
2.50 3.50 3.60
1.80 3.10 3.50
2.40 2.80 3.10
2.70 3.00 3.40
2.30 2.90 3.50
2.80 2.50
3.00 2.10
2.40
2.50
Paired-samples t-test
(differences between dependent groups)
 If we want to compare two variables measured in
the same sample we would customarily use the t-
test for dependent samples.
 For example, we might be interested to see if there

is a significance difference in the mean output per
individual patients before and after an intensive
training (or before and after an intervention).
 In this case, we take a random sample of patients
and record the amount of output each produced
before the training and again after training.
 We then compute the difference between the two

variables (output before training and output after
training) for each case.
 At last, we apply tests to see if the average

difference is significantly different from zero.
Example: The blood pressure (BP) of 30 mothers were
measured before and after taking a new drug. Our aim is to
see if there is a significant difference in the mean blood
pressure.
Subject BP BP Subject BP BP Subject BP BP
before after before after before after
1 130 110 11 125 105 21 120 100
2 125 130 12 120 125 22 115 120
3 140 120 13 135 115 23 130 110
4 150 130 14 145 125 24 140 125
5 120 110 15 115 105 25 110 105
6 130 130 16 125 125 26 120 115
7 120 115 17 115 110 27 105 105
8 135 130 18 130 125 28 125 120
9 140 130 19 135 130 29 130 125
10 130 120 20 125 115 30 120 110
 SPSS output for paired-sample t-test is:
Paired Samples Statistics

Std. Error
Mean N Std. Deviation Mean
Pair 1 BP before 126.833 30 10.379 1.895
drug
BP after drug 118.000 30 9.340 1.705
Paired Differences
Std. Std. Error 95% CI of the Difference Sig. (2-
Mean Deviation Mean Lower Upper t df tailed)
Pair 1 BP before - 8.833 8.272 1.510 5.745 11.922 5.849 29 0.000
BP after
 Since the p-value is less than 0.01, we reject the null

hypothesis. We conclude that there is a significant mean
difference between the mean blood pressure before and after
taking a new drug of women at the 1% level of significant
(taking new drug appreciated, on average).
Examples: The following data shows the SBP levels (mm Hg) in
10 women while not using (baseline) and while using (follow-up)
oral contraceptives. Can we conclude that there is a difference
between mean baseline and follow-up SBP at the 5% level of
significance?
I SBP (baseline) SBP (follow-up)

1 115 128
2 112 115
3 107 106
4 119 128
5 115 122
6 138 145
7 126 132
8 105 109
9 104 102
10 115 117

Chapter 6 Prametric Tests

Uploaded by

Copyright:

Available Formats

Chapter 6 Prametric Tests

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 6 Prametric Tests

Uploaded by

Copyright:

Available Formats

Parametric Tests

Dejen Tesfaw (PhD)

• Thus, the fluctuation (dispersion) in the

Question: Test if there is a significant difference in the mean of distance

• A box plot is a graphical method of detecting

• The box plot of the data under consideration is

• If p-value = 0.03, then we can reject the null hypothesis

• Note: In SPSS output, the p-value

• Thus, we do not reject, H0 : 12  12 and

• The p-value is 0.005 which is less than 5%

• Thus, we reject the hypothesis of equality

• The data is on current value of properties (houses)

• Our aim is to see if there is a significant difference

• To answer this question, we apply pair-wise

• Since the equality of variances assumption is

• The output of Post Hoc Tests (multiple

• The arrangement of counties from the highest mean value of

Group 1 (n = 10) Group 2 (n = 8) Group 3 (n = 6)

 For example, we might be interested to see if there

 We then compute the difference between the two

 At last, we apply tests to see if the average

Paired Samples Statistics

 Since the p-value is less than 0.01, we reject the null

I SBP (baseline) SBP (follow-up)

You might also like